Search

With the growth of web-based applications and the increased popularity of the World Wide Web (WWW), the WWW became the greatest source of information available in the world leading to an increased difficulty of extracting relevant information. Moreover, the content of web sites is constantly changing leading to continual changes in Web users΄ behaviours. Therefore, there is significant interest in analysing web content data to better serve users. Our proposed approach, which is grounded on automatic textual analysis of a web site independently from the usage attempts to define groups of documents dealing with the same topic. Both document clustering and word clustering are well studied problems. However, most existing algorithms cluster documents and words separately but not simultaneously. In this paper, we propose to apply a block clustering algorithm to categorize a web site pages according to their content. We report results of our recent testing of CROKI2 algorithm on a tourist web site. [ABSTRACT FROM AUTHOR]

This paper presents an application of Bayesian networks for network intrusion detection. The described system, named Basset, utilises this technique to enhance the process of misuse-based detection implemented in Snort system. The paper presents the structure of the proposed solution, the role of the Bayesian networks in the detection process and the application of the system to the detection of a real-world attack – an exploitation of a vulnerability in a web browser. [ABSTRACT FROM AUTHOR]

In this paper, we address the problem of semantic hidden errors in Arabic texts. These are spelling errors occurring in valid words and causing semantic irregularities. We first expose the different types of these errors. Then, we present and argue the adopted approach, which is based on the combination of several methods. Next, we describe the context of our work and show the multi-agent architecture of our system. Finally we present the testing framework used to evaluate the implemented system. [ABSTRACT FROM AUTHOR]

An OCR system is presented to understand mathematical formulas in binary printed document images. The system utilizes a novel component-labeling algorithm for extracting local maximum components from image, and uses these components to locate the mathematical formulas. A character recognition algorithm based on neural networks is then adopted. For segmenting merged characters in the image, a novel segmentation algorithm based on a modified SOM neural network was introduced into the system. With the employment of LL(1) grammar, this system can convert the recognition results into a LaTeX file. [ABSTRACT FROM AUTHOR]